An artificial neural network based approach for prediction the thermal conductivity of nanofluids

摘要

Thermal conductivity is an important thermophysical property of nanofluids in many practical heat transfer applications.In this study, a novel approach is proposed to predict the thermal conductivity of nanofluids under multiple operatingparameters. The proposed approach may be extended to be used to other thermophysical properties of nanofluids. TheKohonen’s self-organizing maps (SOM), as an unsupervised artificial neural network (ANN), is used to provide an accurateprediction tool for the problem in hand. Furthermore, SOM, similar to any ANN-based approach, can handle nonlinearand complex input–output relationships with high generalization ability. Comparison of the SOM predicted values withcorresponding available theoretical results as well as experimental data implies high prediction capability of the developedapproach. The proposed approach was utilized to predict thermal conductivity ratio of oxide (Al_2O_3, CuO and TiO_2)/water nanofluids under various operating conditions (nanoparticle size, temperature, and nanoparticle volume fraction).